search

opensafely / booster-effectiveness-autumn-2022

MIT License
1 stars 1 forks source link

Willreview #2

Closed wjchulme closed 1 year ago

elsie-h commented 1 year ago

Thanks!

All looks good, apart from I don't think that we need to detect more AZ doses. I've extracted up to 5 doses of any band here, and joined the "any brand" data to the branded data here.

People with more than 2 doses of AZ are excluded from the study, but we can do that without extracting more than 2 doses of AZ because the brand will be NA for the third dose of AZ in the joined data.

Please confirm you're happy with this and I'll revert the commit.

wjchulme commented 1 year ago

Ah right yes, I understand. I don't think we've ever actually reported a breakdown of vaccine schedules for those that didn't meet the inclusion criteria (eg Pfizer-Pfizer-AZ) so yes, just knowing that the dose is not what we want to include is enough.

So yes, happy for you to revert that bit

elsie-h commented 1 year ago

I made various other changes (meant to make these in a new branch but forgot). Rather than reviewing these in a PR, I think it's easier if I make an issue that highlights the final things that we need to check before running.